Feedback- pipeline usage, --field support

Tobias Girstmair t-passwd at
Thu Aug 15 19:02:48 CEST 2019

The mails in question:

On Thu, Aug 15, 2019 at 04:30:18PM +0000, Robert Ames wrote:
>$ ./src/ ls | grep dream

you can also use `pass find dream` (although it still displays in 
tree-form). patches that output the full path have been proposed 
multiple times already; the consensus was that you shall use
    find ~/.password-store
if you need a complete path. (I personally would like that too, though).
If it were adopted, it needs to be opt-in; --long/-l, --full-path/-f, 
--path/-p have been proposed, if I recall correctly.

>$ ./src/ show rames/ --field=username

definetly useful, but also has been proposed before. passwordstore not 
specifying how fields should be implemented 
( was the reason for it not 
being adopted, as i remember. I'd like that too, but that would require 
abandoning the 'everything goes' approach to data organisation. We'd 
need some robust error handling if it were adopted (what would happen if 
you asked pass(1) to return an non-existing field?  what if the password 
file contained binary data? ...).
My passmenu script currently uses the following to fetch usernames:
    pass show foo | sed -n -e '/fieldname/{s/.*: *//p}'
(yes, that regex isn't fully correct, i know.)

I haven't looked at the code too closely, so I'll refrain from 
commenting on it.


More information about the Password-Store mailing list